Upper zone of growth plate and cartilage matrix associated protein protects cartilage during inflammatory arthritis

Recombinant His-FLAG-tagged Ucma was expressed episomally in HEK293EBNA cells and purified from conditioned medium by affinity chromatography on nickel–nitrilotriacetic acid Sepharose (Qiagen), as reported previously [7]. Collagens type I and II (derived from chicken by pepsin extraction) and recombinant human ADAMTS5 (R&D Systems) were immobilised on PVDF membranes by vacuum blotting. Binding of recombinant Ucma to immobilised proteins was detected as described previously [8]. Briefly, blots were blocked with BSA and incubated in BSA or recombinant Ucma (100 ng/μl). Bound Ucma was detected using rabbit anti-Ucma (UCMA-1; 1:1000) or mouse anti-FLAG (1:1000; Sigma-Aldrich) antibody, anti-rabbit-IgG-HRP or anti-mouse-IgG-HRP antibody, and ECL-based chemoluminiscence.

For solid phase binding assays, ELISA plates were coated with BSA (blanks), collagen I, collagen II or ADAMTS5 (2–250 ng/well, respectively). As reported previously, plates were blocked with BSA and incubated with Ucma (100 ng/ml). Bound Ucma was detected using mouse-anti-FLAG antibody (1:1000), anti-mouse-IgG-HRP antibody and o-phenylenediamine-based colorimetric detection at 492 nm [8]. Incubation with BSA instead of Ucma did not reveal significant signals.

Chondrogenic 4C6 cells were maintained in DMEM/Ham’s F12 medium (Gibco®, ThermoFisher Scientific) with 10% fetal calf serum (FCS), as described previously [12]. For chondrogenic maturation and matrix production, cells were cultured confluently for 7 days in DMEM/Ham’s F12 medium with 10% FCS, 10 mM β-glycerophosphate and 50 μg/ml ascorbate. Subsequently, cells were stimulated with indicated doses of recombinant ADAMTS5, IL-1β or Ucma for 24 h in serum-free DMEM/Ham’s F12 medium before immunohistochemical detection of NITEGE epitopes. The NITEGE-positive chondrocyte culture area was quantified using ImageJ software (n = 3 or 4).

C57BL/6 mice were obtained from Elevage Janvier. We have previously generated Ucma-deficient mice. These mice and their wild-type (WT) littermates were bred in-house under specific pathogen-free conditions. Significant abnormalities in growth, fertility, development or skeletal morphology were not detected in Ucma-deficient mice and they were indistinguishable in size and weight from WT C57/B6 mice [9]. Experimental arthritis in WT and Ucma-deficient mice was induced by transfer of 100 μl K/BxN serum to recipient mice via intraperitoneal (i.p.) injection (serum-induced arthritis (SIA), n = 6–8 per group.). Serum was obtained from 8-week-old K/BxN mice, as reported previously [13, 14]. Ucma treatment of C57BL/6 mice during SIA was performed by daily i.p. injections, each with 100 μl PBS or 2 μg recombinant Ucma in 100 μl PBS, starting at day 1 after serum transfer (n = 3–5 per group). Clinical signs of arthritis, including loss of grip strength (0 = normal grip strength, −1 = mildly reduced, −2 = moderately reduced, −3 = severely reduced and −4 = no grip strength) and paw swelling (0 = no swelling, 1 = detectable, 2 = mild, 3 = moderate and 4 = severe swelling of toes and ankle), were assessed in a blinded manner at the indicated time points, as described previously [14, 15]. Ten or 14 days after serum transfer, the mice were sacrificed. The left hind paw of each mouse was used for serial paraffin sections and the right hind paw was used for micro-computed tomography (micro-CT) or gene expression studies. All experiments were performed with the approval of the local ethics authorities (Government of Mittelfranken, Ansbach, Germany) and according to the regulations of the animal facilities in Germany.

Sections were stained with hematoxylin and eosin (H&E) for the analysis of inflammation, osteophyte formation and bone erosion or with safranin O for the assessment of proteoglycan content of articular cartilage. Osteoclast counts were determined on sections after histochemical detection of tartrate-resistant acid phosphatase (TRAP) using the leukocyte acid phosphatase staining kit (Sigma-Aldrich). All histological quantifications were carried out in a blinded manner by histomorphometry using ImageJ software or an OsteoMeasure system (OsteoMetrics), as reported previously [8, 15]. All analyses were performed on two or three comparable central sections per paw.

Micro-CT image acquisition and analysis of hind paws was performed as reported recently [16]. Briefly, calcified tissue visualisation was performed using the cone-beam Desktop Micro Computer Tomograph “μCT 40” (SCANCO Medical AG, Bruettisellen, Switzerland) at 55 kVp, 145 μA and 200 ms integration time with the operating system “Open VMS” (SCANCO).

Immunohistological detection of Ucma on paw sections from mice was carried out after hyaluronidase-mediated antigen retrieval with UCMA ab-2 antibody (1:500), SignalStain® Boost IHC Detection Reagent (HRP rabbit; Cell Signaling Technologies) and DAB as described recently [8, 9]. Similarly, ADAMTS-generated aggrecan cleavage products containing the NITEGE neoepitope were detected using a polyclonal rabbit anti-NITEGE antibody (Ab1320, 1:1000; IBEX Pharmaceuticals Inc.). Serum levels of Ucma were detected by spotting 10 μl serum from PBS-treated or Ucma-treated mice onto a nitrocellulose membrane. After blocking in 5% BSA, Ucma was detected using UCMA1 antibody (1:1000) and anti-rabbit IgG [9].

RNA in-situ hybridisation on paw sections was carried out with digoxigenin-labelled antisense riboprobes as outlined previously [9, 17]. Briefly, antisense riboprobes for collagen 10 a1 (Col10a1; nucleotides 1689–2114 in NM_009925) and osteocalcin (nucleotides 8–378 in X04142.1) were prepared by in-vitro transcription in the presence of digoxigenin-labelled dNTPs (Roche), as reported previously [17]. Rehydrated paw sections were hybridised with antisense riboprobes at 55 °C overnight. After stringent washing, bound riboprobes were detected with an alkaline phosphatase-conjugated anti-digoxigenin antibody and BM Purple (Roche) detection reagent.

Total RNA was extracted from joint tissue in the paw (n = 3/group), using the RNeasy Fibrous Tissue Mini Kit (Qiagen). cDNA was synthesised using SuperScript II reverse transcriptase (Invitrogen) and mRNA expression relative to cyclophilin A was quantified by real-time RT-PCR using gene-specific primers (Additional file 1), as described previously [7].

Data are presented as the mean ± standard error of the mean. Statistical significance was evaluated by one-way analysis of variance (ANOVA) and Tukey’s multiple comparison test after confirming normal distribution using the Kolmogorov–Smirnov test. Statistical analysis was performed using Graph-Pad Prism software.

We have demonstrated previously that Ucma blocks ADAMTS-specific aggrecanase activity in vitro [8]. In order to identify the mechanism of Ucma-mediated aggrecanase inhibition, we investigated protein–protein interactions between Ucma and ADAMTS5. Therefore, indicated doses of recombinant ADAMTS5 were immobilised on PVDF membranes. Collagen I and collagen II as negative and positive controls, respectively, were also blotted onto the same membranes. Blots were then incubated with recombinant FLAG-tagged Ucma or BSA. Detection of bound Ucma using anti-FLAG (Fig. 1a) or anti-Ucma (Additional file 2) antibody revealed that Ucma directly interacts with ADAMTS5. As expected, Ucma also bound to collagen II but not to collagen I (Fig. 1a) [8]. Solid phase binding assays using ADAMTS5-coated or collagen-coated ELISA plates and Ucma in the liquid phase confirmed binding of Ucma to ADAMTS5 (Fig. 1b). These findings imply that Ucma physically interacts with ADAMTS5 and thereby blocks the aggrecanase activity of ADAMTS5.

In order to test the feasibility of excess Ucma doses to control aggrecanase activities in cartilage under inflammatory conditions, we investigated the effect of recombinant Ucma on ADAMTS-specific aggrecan neoepitope (NITEGE) formation in cultured chondrocytes. Therefore, chondrogenic 4C6 cells were treated with recombinant ADAMTS5 in the absence or presence of Ucma. Afterwards, NITEGE formation in the chondrocyte layer was determined immunohistochemically. As expected, ADAMTS5 induced the extracellular formation of NITEGE-containing aggrecan cleavage products in 4C6 cultures. UCMA, however, inhibited ADAMTS5-induced NITEGE formation in a dose-dependent manner (Fig. 1c, d), confirming Ucma-dependent inhibition of ADAMTS5 aggrecanase. Intriguingly, recombinant Ucma also blocked IL-1β-induced NITEGE formation in 4C6 chondrocyte cultures (Fig. 1c, d). NITEGE formation was hardly detectable in untreated (no ADAMTS5, no IL-1β) 4C6 cultures and Ucma did not affect NITEGE formation under these conditions (Fig. 1c, d). This suggests that excessive Ucma might be feasible for control of pathologic ADAMTS activities in cartilage, which can be found under inflammatory conditions.

This observation prompted us to investigate whether Ucma may not only play a role during degenerative joint diseases, such as OA, but also in inflammatory arthritis. Therefore, we compared Ucma protein expression in hind paws of WT mice with and without serum-induced arthritis (SIA). We observed significant Ucma protein expression in articular cartilage, which, however, was not altered in mice with experimental arthritis (Fig. 2a, b). In accordance, Ucma mRNA levels were not altered during SIA (Additional file 3). Yet, in contrast to healthy controls, mice with SIA exhibited substantial Ucma protein expression at sites of forming osteophytes (Fig. 2c). The main region of Ucma expression during osteophyte formation corresponded to chondrocytes adjacent to Col10a1-expressing hypertrophic chondrocytes (Fig. 2c).

We next investigated whether Ucma may affect the severity of arthritis. Therefore, we studied the progression of SIA in WT and Ucma-deficient mice. To clinically assess the progression of arthritis, we monitored paw swelling and loss of grip strength in a semi-quantitative manner, in controls without induction of arthritis and in WT and Ucma-deficient mice induced for arthritis over 14 days. In contrast to controls, wild-type mice with SIA exhibited joint swelling (Fig. 3a) and loss of grip strength (Fig. 3b). Significant differences in these parameters between WT mice and their Ucma-deficient littermates, however, were not observed. Fourteen days after induction of arthritis, mice were sacrificed and inflamed paw tissue was quantified by histomorphometry. Synovitis was only detected in mice with SIA and its amount did not differ between WT and Ucma-deficient mice (Fig. 3c, d). To study the effect of excess Ucma on experimental arthritis, we systemically administered recombinant Ucma to C57/Bl6 mice with SIA. Systemic administration of Ucma was performed by daily intraperitoneal injections starting 1 day after serum transfer. Clinical signs of arthritis (i.e. paw swelling and loss of grip strength) were determined every other day. Ucma treatment did not affect these disease parameters (Fig. 3e, f). Likewise, histomorphometric quantification of inflammation on paw sections at 10 days after serum transfer did not reveal any effect of Ucma treatment on inflammation (Fig. 3g). Together, these observations indicate that Ucma does not affect joint inflammation.

In order to investigate the impact of Ucma on arthritis-triggered cartilage damage, hind paw sections stained with safranin O were analysed for eroded cartilage surface and proteoglycan loss. As expected, healthy WT or mutant controls did not exhibit any overt cartilage damage in the metatarsal joints (Fig. 4A–C). During SIA, however, Ucma-deficient mice exhibited significantly increased levels of cartilage erosion (Fig. 4A, B) and proteoglycan loss (Fig. 4A, C) compared to WT littermates. This was not associated with significant Ucma-dependent differences in gene expression of ADAMTS4 and ADAMTS5 or MMP3, MMP9 and MMP13 (Additional file 3B–F). However, we observed significantly increased generation of ADAMTS-specific aggrecan degradation products with NITEGE neoepitope in the articular cartilage in arthritic Ucma-deficient mice compared to WT littermates (Fig. 4D, E). This finding indicates an increase in ADAMTS activity in Ucma-deficient mice with SIA, which is in line with Ucma-dependent inhibition of inflammation-triggered ADAMTS aggrecanase activity (Fig. 1) and with the increased proteoglycan loss in Ucma-deficient mice observed by safranin O staining in Fig. 4A.

In order to test, whether Ucma may consequently be a candidate for the treatment of arthritis-triggered cartilage damage in vivo, we investigated SIA-triggered cartilage degradation in C57/Bl6 mice systemically treated with recombinant Ucma. Ten days after serum transfer, Ucma was readily detectable in the serum of treated mice (Fig. 4G). Consistent with our loss-of-function data in Fig. 4A–E, mice treated with recombinant Ucma developed alleviated arthritis-triggered cartilage degeneration during SIA, determined as cartilage erosion (Fig. 4F, H) and proteoglycan loss (Fig. 4F, I) on safranin-O stained paw sections.

Together these data support the notion that Ucma protects from arthritis-triggered ADAMTS-dependent aggrecanolysis and cartilage degradation in SIA and propose Ucma as a candidate for future therapeutic strategies aiming at cartilage health during arthritis.

SIA is characterised by osteophyte formation and since Ucma was found to be expressed in chondrocytes within newly formed osteophytes we were interested in whether Ucma also influences osteophyte formation during arthritis. Therefore, the size of osteophytes at the calcaneus from WT and Ucma-deficient mice with SIA was measured histomorphometrically. Osteophytes in Ucma-deficient mice were substantially smaller than those in their WT littermates (Fig. 5B–D). The decrease in osteophyte size in Ucma-deficient arthritic mice was also detected by micro-CT scans (Fig. 5E). Although the treatment of WT mice with recombinant Ucma did not further increase the size of SIA-induced osteophytes 10 days after induction of arthritis (Additional file 4), these findings support the notion that Ucma promotes osteophyte formation. In contrast, arthritic bone erosion was not significantly affected by Ucma (Fig. 5A).

To investigate the cell type mainly responsible for Ucma-dependent promotion of osteophyte formation, we investigated the cellular composition of hind paw bones from WT and Ucma-deficient mice with and without SIA. Since osteophytes form via a cartilage primordium, we first investigated the presence of hypertrophic chondrocytes by RNA in-situ hybridisation for collagen10a1. The collagen10a1-positive cells were detected at the margins of osteophytes at the calcaneus of mice with SIA. However, there was no apparent difference in hypertrophic chondrocytes between WT mice and their Ucma-deficient littermates (Fig. 6a, b).

Metabolically active osteoblasts were detected by RNA in-situ hybridisation for osteocalcin. As expected, the numbers of osteocalcin-positive osteoblasts considerably increased during SIA at the sites of osteophyte formation. In Ucma-deficient mice, however, substantially lower numbers of osteocalcin-positive osteoblasts were detected. Healthy WT and Ucma-deficient mice, however, did not differ in osteoblast numbers (Fig. 6c, d). In line with the reduction of metabolically active osteoblasts in osteophytes of Ucma-deficient mice, the number of metabolically active osteoclasts within the osteophytes was also significantly reduced (Fig. 6e–g). Hence, staining for TRAP activity showed lower osteoclast numbers in Ucma-deficient mice compared to WT littermates. Receptor-activator of nuclear factor kappa B ligand (Rankl; tumour necrosis factor superfamily member 11 (TNFSF11)) or osteoprotegerin (Opg) mRNA expression, however, was not altered by Ucma (Additional file 3G,H). These data support our earlier findings that pathologic bone remodelling is reduced in Ucma-deficient mice [8].